Scien

Library
Science Cleary
Cleary master's degree major
degree major degree
master's gender
gender female
degree profession
profession

writer writer

Open Domain Question Answering based on

Text Enhanced Knowledge Graph with Hyperedge Infusion

Jiale Han, Bo Cheng, Xu Wang

State Key Laboratory of networking and switching technology,
Beijing University of Posts and Telecommunications
{hanjl,chengbo,wxx}@bupt.edu.cn

Question: Which university did Cleary graduate from?

Abstract
The incompleteness of knowledge base (KB)
is a vital factor limiting the performance of
question answering (QA). This paper proposes
a novel QA method by leveraging text informa-
tion to enhance the incomplete KB. The model
enriches the entity representation through se-
mantic information contained in the text, and
employs graph convolutional networks to up-
date the entity status. Furthermore, to exploit
the latent structural information of text, we
treat the text as hyperedges connecting enti-
ties among it to complement the deficient re-
lations in KB, and hypergraph convolutional
networks are further applied to reason on
the hypergraph-formed text. Extensive ex-
periments on the WebQuestionsSP benchmark
with different KB settings prove the effective-
ness of our model.
1 Introduction
Open domain question answering (QA) is a chal-
lenging task that requires answering the factual
questions in natural language. According to the
structure of supporting information, QA system
can be divided into knowledge-based QA (KBQA)
(Bordes et al., 2015; Zhang et al., 2018) and text-
based QA (TBQA) (Welbl et al., 2018; Yang et al.,
2018). KBQA obtains the answers by a structured
knowledge base, which is easy to query and reason
with but limited by the incompleteness of well-
designed triples. TBQA’s supporting information
is plain text containing rich semantic and latent
structural information, however, it’s difficult for a
machine to understand. The complementary prop-
erties inspire us to fuse these two kinds of data to
enhance the incomplete KB and further improve
the QA system’s performance.
Some work has already been proposed. Das
et al. (2017) represent KB and text using universal
schema and apply memory networks, but lack the
Text: Cleary obtained her master's
degree in Library Science at the
University of Washington. University of Library
Washington Science
Cleary Cleary
master's degree major Library master's
degree major
degree Science
degree
gender
profession gender
profession
female
writer
writer female
Figure 1: Example of a question with its related KB
and text. The KB is incomplete to answer the question,
which lacks relation ”graduated university” and entity
”University of Washington”. By completing the miss-
ing information with adding text as hyperedge, we can
handle the question more effectively.
association between KB and text. Sun et al. (2018)
build a heterogeneous graph with entities and text
as nodes and employ a graph based method. Xiong
et al. (2019) first encode entities in KB by graph at-
tention networks and then read text with the help of
accumulated entity knowledge. Although good re-
sults have been achieved, the text information is not
fully utilized, especially the relation information
among the entities contained in the text. Figure 1
shows an example that the KB is insufficient to an-
swer the question. This question can be adequately
answered by using the structural information of the
text to bring high-level relationships.
In this paper, we propose a novel QA model
based on text enhanced knowledge graph, which
enriches entity representation by text semantic in-
formation and complements the relations in KB
through structural information of the text. Specifi-
cally, the model firstly encodes entities in KB com-
bining text information and applies graph convo-
lutional networks (GCN) (Wu et al., 2020) to rea-
son across KB. Note that a document usually men-
tions multiple entities, we convert the unstructured
text into a structured hypergraph by regarding text

1475
Findings of the Association for Computational Linguistics: EMNLP 2020, pages 1475–1481
November 16 - 20, 2020.
2020
c Association for Computational Linguistics
as hyperedge connecting entities among text, and
then employ hypergraph convolutional networks
(HGCN) (Feng et al., 2019; Yadati et al., 2019) to
further update the entity states. Finally, the model
predicts the final answers.
Our highlights are summarized as follows: 1) We
novelly treat documents as high-order relations (hy-
peredges) connecting entities mentioned in them.
2) We apply Hypergraph Convolutional Networks
to reason and propose the dual-step attention to
catch the importance of different entities and doc-
uments. 3) Extensive experiments conducted on
the widely used WebQuestionsSP (Yih et al., 2016)
with different KB settings demonstrate our model
is effective.
2 Related Work
The combination of knowledge base and text in
QA is a challenging task, which has attracted many
researchers’ attention. The work of (Das et al.,
2017) extends universal schema to question an-
swering and employs Key-Value Memory networks
to process to text and KB. Sun et al. (2018) re-
gard documents as heterogeneous nodes and com-
bine them with entities in KB to form a uniform
graph. The model proposed by Xiong et al. (2019)
contains a graph-attention based KB reader and a
knowledge-aware text reader. Some other work
focuses on retrieving a small graph that contains
just the question-related information (Sun et al.,
2019) and the interpretability of QA on KB and
text (Sydorova et al., 2019). These methods lack
considering the high-order relationship among the
entities contained in the text. This paper regards the
text as hyperedge and further employs hypergraph
convolutional networks.
Hypergraph convolutional networks (Feng et al.,
2019; Yadati et al., 2019) utilize hypergraph struc-
ture rather than a general graph to represent the
high-order correlation among data entirely, and
hypergraph attention (Bai et al., 2019) further en-
hances the ability of representation learning by us-
ing an attention module.
3 Model
3.1 Task Definition
To maintain consistency and fairness, we adopt
the same setting as Sun et al. (2018) that builds
a subgraph for each question. Specifically, given
a question q = (w1 , w2 , ..., w|q| ), the related sub
knowledge graph K = (V, E, T ) is extracted by the
1476
Answer Prediction
HGCN
GCN
candidates documents
Input Encoder
KB query Text
Figure 2: The overview of the model. We utilize the se-
mantic information mentioned in the text to enrich the
entity representation, and novelly treat text as hyper-
edges to complement the relation in incomplete KB.
Personalized PageRank (Haveliwala, 2002), where
V is the entity set, E is the relation set, and T con-
tains a set of triples (vh , r, vt ) indicated there is
a relation r ∈ E between vh ∈ V and vt ∈ V.
Also a relevant text corpus D = d1 , d2 , ..., d|D|
is retrieved from Wikipedia by an off-the-shelf
document retriever (Chen et al., 2017), which
di = (w1 , w2 , ..., w|di | ) represents a document and
the entities mentioned in documents have been
linked. The task requires to extract answers from
all KB and document entities. The overview of our
model is shown in Figure 2.
3.2 Input Encoder
Query and Text Encoder: Let Xq ∈ R|q|×n and
Xd ∈ R|d|×n be the embedding matrices of query
q and document d ∈ D, where n is the embed-
ding dimension. Bi-LSTM networks (Hochreiter
and Schmidhuber, 1997) are applied to encode the
query and document separately and get the hidden
states Hq ∈ R|q|×h and Hd ∈ R|d|×h , h is the hid-
den dimension of bi-LSTM. Then we compute the
representation of query hq and document hd with
attention mechanism.
hq = HT
q softmax(fq (Hq )) ∈ R
h×1
hd = HT T
d softmax(fd (Hd Hq )) ∈ R
h×1
where T represents matrix transposition, fq is a
linear network which converts h dimension to 1 αi = σ(hT (l1 )
q fa ([hvi ; hri ]))
dimension, and fd converts |q| dimension to 1 di-
where W1 ∈ Rh×h , W2 ∈ Rh×2h are learnable
mension.
parameters, Nv represents the adjacent triple set of
KB Encoder: Each entity v ∈ V is initialized entity v, fa converts 2h dimension to h dimension,
by pre-trained knowledge graph embedding xv ∈ l1 represents the current GCN layer, which has a
Rn×1 . And relation is initialized by semantic vec- total of L1 layers, and σ is the sigmoid function.
tor and KG embedding. Specifically, for relation
HGCN for Hypergraph-Formed Text: The
r ∈ E and its KG embedding xr ∈ Rn×1 , we to-
model regards plain text as hyperedges connect-
kenize it as r = (w1 , w2 , ..., w|r| ) and feed into
ing the entities among the text to complement the
bi-LSTM layer with word embedding to get the
lack of relations in KB. HGCN is employed to en-
hidden states Hr ∈ R|r|×h , then calculate the rep-
code on the hypergraph-formed text. What’s more,
resentation hr as follows.
dual-step attention catches the importance of dif-
|r|×h ferent entities and documents. Formally, at layer
Hrq = softmax(Hr HT
q )Hq ∈ R
l2 , the model first transfers the entity feature to
0 the connected hyperedges to form the document
Hr = [Hr ; Hrq ] ∈ R|r|×2h
representation,
0 0 0
hr = HrT softmax(fr1 (Hr )) ∈ R2h×1 0 (l +1) 0 (l ) X 0
0
hd 2 = W3 hd 2 + βi W4 hv(li 2 ) ∈ Rh×1
hr = fr2 ([hr ; xr ]) ∈ Rh×1 vi ∈Nd

where [; ] denotes column-wise concatenation, fr1 βi = σ(hT

0
(l2 )
q hvi )
is a linear network which converts 2h dimension to
1 dimension, and fr2 converts 2h + n dimension where W3 , W4 ∈ Rh×h are learnable parame-
0 (0) (L ) 0 (0)
to h dimension. ters, hv = hv 1 , hd = hd , and Nd represents
the connected entity set of document d. Then the
3.3 Reasoning over Text Enhanced model gathers the documents’ information to up-
Knowledge Graph date the connected entity states.
This component utilizes text information to im- 0 0
X 0 (l +1)

prove the incomplete KB by enriching entity repre- hv(l2 +1) = W5 hv(l2 ) + γi W6 hdi 2 ∈ Rh×1
di ∈Dv
sentation and adding hyperedges, and applies GCN
and HGCN to reasoning. 0 (l +1)
γi = σ(hT
q hdi
2
)
GCN for Entity-Enriched KB: To utilize the
where W5 , W6 ∈ Rh×h are learnable parameters.
rich semantic information contained in the text,
we construct a binary matrix M, where Mvd ∈ 3.4 Answer Prediction
R|d|×1 indicates the span of entity v in document After L1 GCN layers and L2 HGCN layers, the
d, and pass information from documents to entities model finally predicts the probability of each entity
0
to form text-aware entity representation xv , then being the answer,
(0)
concatenate with xv as initial node state hv . 0
pv = σ(fout (hv(L2 ) ))
0
X
xv = HT v
d Md ∈ R
h×1

d∈Dv
where fout converts h dimension to 1 dimension.

0 4 Experiments
h(0)
v = fv ([xv ; xv ]) ∈ R
h×1

4.1 Dataset
where Dv is the linked documents set of entity
v, fv converts h + n dimension to h dimension. WebQuestionsSP (Yih et al., 2016) is a multi-
Then the model learns the entity representation by answer QA dataset which contains 4737 questions.
aggregating the connected entity feature. In our experiments we adopt the dataset 1 prepro-
cessed by Sun et al. (2018). Table 1 shows the
X 1
h(l
v
1 +1)
= W1 h(l
v
1)
+ αi W2 [h(l 1)
vi ; hri ] ∈ R
h×1 https://github.com/OceanskySun/
(vi ,ri )∈Nv
GraftNet

1477
 questions avg candidates avg linked avg entities
Dataset avg answers
train / dev / test KB / Text / KB+Text documents in documents
WebQSP 2848 / 250 / 1639 384.6 / 141.6 / 515.1 43.6 11.2 4.6

Table 1: Statistical results of the dataset WebQuestionsSP.

Model KB-only Text-only KB+Text The GCN layer L1 and HGCN layer L2 are 1 and
KVMem 46.7 / 38.6 23.2 / 13.0 40.5 / 30.9 2 separately. The average runtime for one epoch is
GraftNet 66.7 / 62.4 25.3 / 15.3 67.8 / 60.4 5 minutes, and we set the max number of epochs to
SG-KA 66.5 / 58.0 -/- 67.2 / 57.3 200. The number of parameters is 69 million. The
PullNet 68.1 / - 24.8 / - -/- Adam optimizer (Kingma and Ba, 2015) is applied
Ours 66.9 / 60.1 27.2 / 17.1 68.4 / 60.6 to minimize the binary cross-entropy loss with a
learning rate of 0.0005. The threshold for F1 is set
Table 2: Hits@1 / F1 scores on WebQSP.
to 0.05.

statistics of the dataset and retrieved subgraphs for
the questions, including KB and linked text. In par-
ticular, the average number of linked entities in the
documents is 4.6, which illustrates the rationality
of adopting hyperedges.
4.2 Baseline Methods
We compare our methods with the following mod-
els:
• KVMemNet (Miller et al., 2016) is an end-to-
end memory network which stores KB facts
and text into key-value pairs.
• GraftNet (Sun et al., 2018) combines KB and
text with the early fusion strategy and applies
a graph-based model.
• SG-KA Reader (Xiong et al., 2019) proposes
two components to reason over KB and incor-
porate entity information to text.
• PullNet (Sun et al., 2019) is a QA framework
for learning how to retrieve small sub-graph
related to answering the question.
4.3 Training Details
The model is implemented in PyTorch (Paszke
et al., 2019) and trained on one Nvidia Tesla P40
GPU. We apply 100-dimensional TransE embed-
dings (Bordes et al., 2013) for entities and relations,
and 300-dimensional GloVe embeddings (Penning-
ton et al., 2014) for question and text words. The
word numbers of questions and documents are lim-
ited to be 10 and 50. The hidden size is set to 100.
We select the hyperparameter values by manual
tuning to perform the best results on the validation
dataset. The dropout is 0.2, and the batch size is 8.
4.4 Results
Main Results: The metrics adopted in the exper-
iments are Hits@1, which is the accuracy of the top
answer predicted by the model, and F1, which rep-
resents the ability to predict all answers. As shown
in table 2, we experiment with our model under KB-
only, Text-only, and KB+Text settings and compare
them with baseline methods. Our model gets com-
petitive performance in the KB-only setting and
achieves the best results in the other two settings,
especially in the Text-only setting, Hits@1 and F1
are 1.9% and 1.8% higher than the second-best
method respectively, which shows the validity of
treating documents as hyperedges. The promising
performance may inspire us to handle similar tasks
that build plain text to hypergraph and apply ef-
ficient HGCN. In KB+Text’s setting, our method
also achieves the best performance, proving that
our proposed enhancement strategy can effectively
enhance incomplete KB by fully introducing the
semantic and structural information implied in the
text. In particular, our model improves a lot com-
pared with KB-only, more than the work of (Sun
et al., 2018), which demonstrates our way that treat-
ing documents as hyperedges is more productive
than regarding them as heterogeneous nodes.
Different KB Setting: Following the work of
Sun et al. (2018) that the KB is downsampled to
different extents, we experiment on 10%, 30%, and
50% KB settings, which represents the percentage
of required evidence covered by KB to simulate
the situation of incomplete KB, and analyze the im-
pact of the text on model performance. As shown
in table 3, our model obtains the promising per-
formance in the KB-only setting, especially the
F1 metric all achieves the highest values, demon-

1478
 10% 30% 50%
Model
KB-only KB+Text KB-only KB+Text KB-only KB+Text
KVMem 12.5 / 4.3 24.6 / 14.4 25.8 / 13.8 27.0 / 17.7 33.3 / 21.3 32.5 / 23.6
GraftNet 15.5 / 6.5 31.5 / 17.7 34.9 / 20.4 40.7 / 25.2 47.7 / 34.3 49.9 / 34.7
SG-KA 17.1 / 7.0 33.6 / 18.9 35.9 / 20.2 42.6 / 27.1 49.2 / 33.5 52.7 / 36.1
PullNet -/- -/- -/- -/- 50.3 / - 51.9 / -
Ours 18.3 / 7.9 33.7 / 19.9 35.2 / 21.0 42.8 / 27.5 49.3 / 34.3 52.8 / 37.1

Table 3: Hits@1 / F1 scores under different KB settings.

10%KB+Text
 .90HP Model
*UDIW1HW Hits@1 F1
6*.$ Full Model 33.7 19.9
 2XUV
−GCN attention 33.3 19.3
+LWV#

 −dual-step attention 32.5 18.9



  
.%IUDFWLRQ
Figure 3: Improvement of KB+Text over KB-only un-
der different KB fraction setting.
strating the ability of our method for multi-answer
prediction. After combining the text, our model
achieves the best results compared with the base-
line methods. What’s more, the performance in the
KB+Text setting has been significantly improved
over the KB-only setting. The more incomplete the
KB, the more obvious the performance improve-
ment, which shows it can effectively use the docu-
ment information to complete and enhance the KB,
so as to further improve the performance of QA.
In order to intuitively visualize the improvement,
figure 3 displays the increment of all models after
adding text under different settings (KB+Text−KB-
only). We can observe that our method achieves
the largest or almost the largest increment. What’s
more, we notice the text information improves the
performance obviously in the case of incomplete
KB, but may cause the extra interference when the
KB is sufficient to support answering questions,
which even lead to performance degradation. This
makes us think about how to effectively use text
to further improve the performance of question an-
swering under the full KB setting.
Ablation Study: An ablation study is conducted
to evaluate the benefits of different components in
−entity-enriched KB 32.8 18.7
Table 4: Experimental results of ablation study.
 the model. Table 4 shows the results under 10%
KB setting. From the second and third rows, the
attention mechanism adopted by the model is effec-
tive, especially the dual-step attention proposed at
the HGCN layer, which brings 1.2% improvement
of Hits@1. The strategy of entity-enriched KB also
increases Hits@1 by 0.9%, proving its validity.
5 Conclusion
We propose a QA method that aims to enhance the
incomplete KB by text information, which fully
explored the semantic and latent structural infor-
mation in the text. In particular, the text is treated
as hyperedges to complement the lack of relations
in KB. The model first applies GCN to encode the
entity-enriched KB, then employs HGCN to further
reason over hypergraph-formed text, and predicts
the final answers. Experimental results on the We-
bQuestionsSP benchmark prove the effectiveness
of our model and each component.
Acknowledgements
This work is supported by the Fundamental Re-
search Funds for the Central Universities under
grant 2020XD-A07-1, the National Key Research
and Development Program of China under grant
2018YFB1003804, in part by the National Nat-
ural Science Foundation of China under grant
61921003, 61972043. Jiale Han is supported by
BUPT Excellent Ph.D. Students Foundation (Grant
No. CX2020102). We would like to thank all re-
viewers for careful and valuable comments.

1479
References
Song Bai, Feihu Zhang, and Philip H. S. Torr. 2019.
Hypergraph convolution and hypergraph attention.
CoRR, abs/1901.08150.
Antoine Bordes, Nicolas Usunier, Sumit Chopra, and
Jason Weston. 2015. Large-scale simple ques-
tion answering with memory networks. CoRR,
abs/1506.02075.
Antoine Bordes, Nicolas Usunier, Alberto Garcı́a-
Durán, Jason Weston, and Oksana Yakhnenko.
2013. Translating embeddings for modeling multi-
relational data. In Advances in Neural Information
Processing Systems 26: 27th Annual Conference on
Neural Information Processing Systems 2013. Pro-
ceedings of a meeting held December 5-8, 2013,
Lake Tahoe, Nevada, United States, pages 2787–
2795.
Danqi Chen, Adam Fisch, Jason Weston, and Antoine
Bordes. 2017. Reading wikipedia to answer open-
domain questions. In Proceedings of the 55th An-
nual Meeting of the Association for Computational
Linguistics, ACL 2017, Vancouver, Canada, July 30 -
August 4, Volume 1: Long Papers, pages 1870–1879.
Association for Computational Linguistics.
Rajarshi Das, Manzil Zaheer, Siva Reddy, and Andrew
McCallum. 2017. Question answering on knowl-
edge bases and text using universal schema and
memory networks. In Proceedings of the 55th An-
nual Meeting of the Association for Computational
Linguistics, ACL 2017, Vancouver, Canada, July 30
- August 4, Volume 2: Short Papers, pages 358–365.
Association for Computational Linguistics.
Yifan Feng, Haoxuan You, Zizhao Zhang, Rongrong Ji,
and Yue Gao. 2019. Hypergraph neural networks.
In The Thirty-Third AAAI Conference on Artificial
Intelligence, AAAI 2019, The Thirty-First Innova-
tive Applications of Artificial Intelligence Confer-
ence, IAAI 2019, The Ninth AAAI Symposium on Ed-
ucational Advances in Artificial Intelligence, EAAI
2019, Honolulu, Hawaii, USA, January 27 - Febru-
ary 1, 2019, pages 3558–3565. AAAI Press.
Taher H. Haveliwala. 2002. Topic-sensitive pagerank.
In Proceedings of the Eleventh International World
Wide Web Conference, WWW 2002, May 7-11, 2002,
Honolulu, Hawaii, USA, pages 517–526. ACM.
Sepp Hochreiter and Jürgen Schmidhuber. 1997.
Long short-term memory. Neural Computation,
9(8):1735–1780.
Diederik P. Kingma and Jimmy Ba. 2015. Adam: A
method for stochastic optimization. In 3rd Inter-
national Conference on Learning Representations,
ICLR 2015, San Diego, CA, USA, May 7-9, 2015,
Conference Track Proceedings.
Alexander H. Miller, Adam Fisch, Jesse Dodge, Amir-
Hossein Karimi, Antoine Bordes, and Jason We-
ston. 2016. Key-value memory networks for di-
rectly reading documents. In Proceedings of the
1480
2016 Conference on Empirical Methods in Natural
Language Processing, EMNLP 2016, Austin, Texas,
USA, November 1-4, 2016, pages 1400–1409. The
Association for Computational Linguistics.
Adam Paszke, Sam Gross, Francisco Massa, Adam
Lerer, James Bradbury, Gregory Chanan, Trevor
Killeen, Zeming Lin, Natalia Gimelshein, Luca
Antiga, Alban Desmaison, Andreas Köpf, Edward
Yang, Zachary DeVito, Martin Raison, Alykhan Te-
jani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang,
Junjie Bai, and Soumith Chintala. 2019. Pytorch:
An imperative style, high-performance deep learn-
ing library. In Advances in Neural Information Pro-
cessing Systems 32: Annual Conference on Neu-
ral Information Processing Systems 2019, NeurIPS
2019, 8-14 December 2019, Vancouver, BC, Canada,
pages 8024–8035.
Jeffrey Pennington, Richard Socher, and Christopher D.
Manning. 2014. Glove: Global vectors for word
representation. In Proceedings of the 2014 Confer-
ence on Empirical Methods in Natural Language
Processing, EMNLP 2014, October 25-29, 2014,
Doha, Qatar, A meeting of SIGDAT, a Special Inter-
est Group of the ACL, pages 1532–1543. ACL.
Haitian Sun, Tania Bedrax-Weiss, and William W. Co-
hen. 2019. Pullnet: Open domain question answer-
ing with iterative retrieval on knowledge bases and
text. In Proceedings of the 2019 Conference on
Empirical Methods in Natural Language Processing
and the 9th International Joint Conference on Nat-
ural Language Processing, EMNLP-IJCNLP 2019,
Hong Kong, China, November 3-7, 2019, pages
2380–2390. Association for Computational Linguis-
tics.
Haitian Sun, Bhuwan Dhingra, Manzil Zaheer, Kathryn
Mazaitis, Ruslan Salakhutdinov, and William W. Co-
hen. 2018. Open domain question answering using
early fusion of knowledge bases and text. In Pro-
ceedings of the 2018 Conference on Empirical Meth-
ods in Natural Language Processing, Brussels, Bel-
gium, October 31 - November 4, 2018, pages 4231–
4242. Association for Computational Linguistics.
Alona Sydorova, Nina Pörner, and Benjamin Roth.
2019. Interpretable question answering on knowl-
edge bases and text. In Proceedings of the 57th Con-
ference of the Association for Computational Lin-
guistics, ACL 2019, Florence, Italy, July 28- August
2, 2019, Volume 1: Long Papers, pages 4943–4951.
Association for Computational Linguistics.
Johannes Welbl, Pontus Stenetorp, and Sebastian
Riedel. 2018. Constructing datasets for multi-hop
reading comprehension across documents. Trans.
Assoc. Comput. Linguistics, 6:287–302.
Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong
Long, Chengqi Zhang, and Philip S. Yu. 2020. A
comprehensive survey on graph neural networks.
IEEE Transactions on Neural Networks and Learn-
ing Systems, pages 1–21.
Wenhan Xiong, Mo Yu, Shiyu Chang, Xiaoxiao Guo,
and William Yang Wang. 2019. Improving ques-
tion answering over incomplete kbs with knowledge-
aware reader. In Proceedings of the 57th Confer-
ence of the Association for Computational Linguis-
tics, ACL 2019, Florence, Italy, July 28- August 2,
2019, Volume 1: Long Papers, pages 4258–4264. As-
sociation for Computational Linguistics.
Naganand Yadati, Madhav Nimishakavi, Prateek Ya-
dav, Vikram Nitin, Anand Louis, and Partha P. Taluk-
dar. 2019. Hypergcn: A new method for train-
ing graph convolutional networks on hypergraphs.
In Advances in Neural Information Processing Sys-
tems 32: Annual Conference on Neural Information
Processing Systems 2019, NeurIPS 2019, 8-14 De-
cember 2019, Vancouver, BC, Canada, pages 1509–
1520.
Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Ben-
gio, William W. Cohen, Ruslan Salakhutdinov, and
Christopher D. Manning. 2018. Hotpotqa: A dataset
for diverse, explainable multi-hop question answer-
ing. In Proceedings of the 2018 Conference on
Empirical Methods in Natural Language Process-
ing, Brussels, Belgium, October 31 - November 4,
2018, pages 2369–2380. Association for Computa-
tional Linguistics.
Wen-tau Yih, Matthew Richardson, Christopher Meek,
Ming-Wei Chang, and Jina Suh. 2016. The value of
semantic parse labeling for knowledge base question
answering. In Proceedings of the 54th Annual Meet-
ing of the Association for Computational Linguistics,
ACL 2016, August 7-12, 2016, Berlin, Germany, Vol-
ume 2: Short Papers. The Association for Computer
Linguistics.

Yuyu Zhang, Hanjun Dai, Zornitsa Kozareva, Alexan-

der J. Smola, and Le Song. 2018. Variational
reasoning for question answering with knowledge
graph. In Proceedings of the Thirty-Second AAAI
Conference on Artificial Intelligence, (AAAI-18),
the 30th innovative Applications of Artificial Intel-
ligence (IAAI-18), and the 8th AAAI Symposium
on Educational Advances in Artificial Intelligence
(EAAI-18), New Orleans, Louisiana, USA, February
2-7, 2018, pages 6069–6076. AAAI Press.

1481